Rubber Toughened Thermosets - Advances in Chemistry (ACS

27 Rubber Toughened Thermosets C. K. RIEW, E. H. ROWE, and A. R. SIEBERT

Downloaded by EAST CAROLINA UNIV on August 28, 2013 | http://pubs.acs.org Publication Date: June 1, 1976 | doi: 10.1021/ba-1976-0154.ch027

B. F. Goodrich Co., Research and Development Center, Brecksville, Ohio 44141

Epoxy resins are toughened with functionally terminated liquid polymers. Small rubbery domains of a definite size and shape are formed in situ during cure. The selectivity and reactivity of the functional groups of liquid polymers are critical for the in situ formation of rubber particles. These systems show marked increases in impact and crack resistance. Shear deformations are dominant in resins toughened with small particles (< 0.5 μm). Crazing is associated with microvoid development around the rubber particles in resins toughened with large particles (1-5 μm). Maximum toughness is obtained under conditions of combined shear and craze deformations. This condition is obtainable when there is bimodal distribution of particle sizes in the rubber toughened resins. Criteria for rubber toughening of brittle resins and failure mechanisms are discussed. These criteria are based on chemical and morphological observations. A small amount of rubber in the form of discrete particles in a glassy thermoset resin, such as epoxy resin, can greatly improve its crack resistance and impact strength. This improvement is achieved without significantly decreasing thermal or mechanical properties. McGarry with his co-workers (1,2) and Rowe et ah (3), emphasized using Hycar C T B N (liquid carboxyl terminated butadiene-acrylonitrile copolymer) as a toughener for epoxy resins. They determined the effects of composition, molecular weight of the C T B N , and particle size of the rubber phase on fracture energy improvement in epoxy resin systems. Siebert and Riew (4) described the chemistry of the in situ particle formation. They proposed that the composition of the particle is a mixture of linear CTBN-epoxy copolymers and crosslinked epoxy resin. The polymer morphology of the C T B N toughened epoxy systems was investigated by Rowe (5) using transmission electron microscopy by carbon replication of fracture surfaces. Riew and Smith (6) supported the 326 In Toughness and Brittleness of Plastics; Deanin, R., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1976.

27.

RBEW E T A L .

327

Rubber Toughened Thermosets


character of the discrete rubbery phase using a new osmium staining technique. Rowe and Riew (7) suggested deformation mechanisms based on morphology. They observed shear bands and/or crazes depending on the particle sizes in C T B N toughened epoxy systems. The experiments described above were confined to C T B N toughened epoxy resins. However we believe that the technology should apply to thermoplastics or other thermosets. In this paper we show how the chemistry, physics, and morphology of the C T B N toughened epoxy resins relate to the thermal-mechanical properties of these systems. Experimental Materials. Hycar C T B N is a registered trade name of a carboxylterminated, liquid copolymer of butadiene and acrylonitrile (B. F. Goodrich Chemical Co.). For most purposes it can be represented structurally as: HOOC—[(CH —CH=CH—CH ) 2

2

—(CH —CH) 2

Y

] —COOH M

C N

where averaged x = 5, y = 1, and m= 10. Cure agent D is 2-ethylhexanoic acid salt of DMP-30 (Shell Chemical Co.). DMP-30 is tris-(N,Ndimethylaminomethylene) phenol (Rohm and Haas). Piperidine was used as received (Fisher Scientific Co.). Epon 828 is the liquid diglycidyl ether of bisphenol-A resin ( D G E B A ) (Shell Chemical Co.). Bisphenol A is the trade name of 4,4'-isopropylidene diphenol. It has a melting point of 1 5 2 ° - 1 5 3 ° C and a molecular weight of 228.3 (Dow Chemical Co.). Procedures. P R E P A R A T I O N O F C A S T D O U B L E - C A N T I L E V E R TEST SPECIMENS.

CLEAVAGE

C A S T S A M P L E . This material consisted of 100 parts D G E B A resin (Epon 828), five parts Hycar C T B N , and five parts piperidine (or DMP30). For the following procedure we used a 15 X 25 cm tray. Proceed as follows. Place Epon 828 (285.7 g, ca. 0.76 mole) and the Hycar C T B N (14.3 g, ca. 0.04 mole) in a 500-ml beaker. Stir the mixture mechanically for at least 5 min at room temperature or at an elevated temperature. The mixture is usually turbid with normal Hycar C T B N (acrylonitrile content, 17-18%). Add piperidine (14.3 g, 0.17 mole) in one portion with stirring. Continue the stirring another 5 min. The mixture becomes clear. Pour the mixture inta a Teflon-coated metal tray which has been leveled and preheated at 120 °C in an oven. The mixture has a working life of at least 2 hr. Cure at 120 °C for 16 hr. Cool the sample gradually to avoid warpage, and remove it from the tray. Cut into three double cantilever specimens (3.8 X 23 cm) for fracture energy test. H Y C A R C T B N - B I S P H E N O L A - E P O N 828 C A S T S A M P L E S .

A standard

recipe follows: 100 parts D G E B A resins (Epon 828), 24 parts bisphenol A, five parts liquid rubber ( C T B N ) , and five parts piperidene. It is often desirable to have a stock mixture of bisphenol A and Epon 828. Since the shelf stability of the mixture is poor at room temperature, it should be refrigerated. Melt the bisphenol A in Epon 828 with stirring

In Toughness and Brittleness of Plastics; Deanin, R., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1976.

328

TOUGHNESS A N D BRITTLENESS O F PLASTICS

at 1 2 0 ° C using an oil bath for ca. 2 0 min. Use essentially the same procedure to prepare cast samples according to the preceding section. FRACTURE

SURFACE

ENERGY

C A L C U L A T I O N : BERRY'S

METHOD

(8).

The dimensions of double cantilever test specimen and techniques are illustrated in Ref. 3. We determined the surface work by substituting the measured quantities from the cleavage tests into Griffith's equation. Three cantilever bars were used. The fracture surface work, y(hj/m ), is given by: 2

_


7

"

n n 4 ' wl

where F = applied force, kgf; 8 = deflection of one cantilever, m; w = crack width, m; I = crack length, m; n = experimental constant. T E N S I L E TESTS. After a sample has been tested for fracture energy, the cleaved bars are used for further mechanical testing. Three of the bars are cut into tensile specimens according to ASTM D - 6 3 8 test methods. G A R D N E R I M P A C T . A single bar remaining from the cleavage test is used for impact resistance. We used a Gardner impact tester and determined an average weight times drop height for a dart of 1.27 cm radius. H E A T DISTORTION T E S T . T W O remaining bars from the cleavage tests are used to determine heat distortion temperature in accordance with ASTM D - 6 2 1 test methods. These thermal-mechanical tests are necessary to determine whether a brittle resin has been truly toughened or whether it has been merely flexibilized. The morphology of the resin also effectively describes a true toughening situation and can aid immeasurably in explaining departures from true toughening. E L E C T R O N MICROSCOPY. We replicated the surfaces developed from the double cantilever cleavage test and the surface developed from the tensile test specimen. Replicas are first taken with a gelatin solution from the fracture surface (9). A typical electron micrograph of a CTBN-epoxy system is shown in Figure 1 . This micrograph is from a surface of a fractured cantilever cleavage specimen. It shows a uniformly dispersed rubber phase in a brittle epoxy matrix. Results and Discussion Epoxy resins toughened with reactive liquid polymers consist of a continuous rigid epoxy resin phase and a dispersed rubber phase. The rubber phase is formed in situ during the early stage of the cure reaction. We do not know the precise chemical composition of the particles. We propose that the particle composition is a mixture of linear copolymers of CTBN-epoxy resin and homopolymers of epoxy resin (4). The rubbery domains impart some ductility which in turn improves the toughness or crack resistance by absorbing strain energy. We have demonstrated a technique by which a two-phase thermoset system is produced to form a toughened system. Also we will describe how the toughening mechanism is related to the chemistry and morphology of these toughened epoxy resin systems.



27.

RIEW E T A L .

Figure I.


329

Typical electron micrograph of a CTBN-toughened epoxy resin (X4400)

Effect of the Selectivity and Reactivity of Functional Groups. The acidic functional end groups such as carboxylic, phenolic, hydroxylic, etc. are highly selective in forming anions with amine catalysts. The order of selectivity is (10): O R — C

>

R—@M3H

>

R__CH —OH > R—CH —SH 2

2

\)H The order of reactivity of functional groups (anions) toward the epoxy group is: O R—CH —S 2

e

>

R—CH —O 2

e

>

R

—

O

e

>

R — /

The polymerization in the discrete phase is an anionic addition between the difunctional liquid rubber and the epoxy resin which gives a linear block copolymer. Thus the selectivity and reactivity of the end groups of the liquid rubber are critical. Table I shows the fracture energies for epoxy resin which have been toughened by liquid polymers with different functional groups. Clearly the selectivity of the functional groups is more important than the reactivity. M T B N , with the most reactive mercaptan end groups, had the least improvement in fracture energy; the cured material is trans-


330

TOUGHNESS AND BRITTLENESS OF PLASTICS

Table I.

Effect of Functional End Groups on Fracture Energy

Liquid AcrylonitrileButadiene Copolymer

Type of Functional End Groups

Hycar C T B N


PTBN

m0>~

phenol; —

6

oh

ETBN"

epoxy;

HTBN MTBN

hydroxyl; -- R — C H O H mercaptan; — R — C H / S H

6

6

--- R — C H — C H

V 2

2

2

Second Phase

Fracture. Energy, hJ/m

Rubbery Domain

Domain Size, [im

28

yes

1-5

26-30

yes

1-5

18-26

yes

2 tt2

of-COOH 1 Terminal Pendent

— no no yes yes

— no yes yes no

Fracture Energy hJ/m

Rubbery Domain

1.75 3.5 7.9 17.5 28.0

no yes yes yes yes

2

Epon 828 (100 parts), liquid rubber (5 parts), and piperidine (5 parts); cured at 120°C for 16 hr. 0



27.

RIEW E T A L .

331


We believe that the Hycar 1312 is not cured. In these systems there exist only the van der Waals attractive and/or hydrogen bonding type forces between the epoxy matrix and the liquid rubber particles. In such two-phase systems the liquid rubber merely acts as a diluent and will show physical properties analogous to mechanical mixtures of two different polymers. Table II shows that the chemical bonding between the epoxy matrix and the rubbery phase is important. The terminal reactive groups are more effective than the pendant groups in toughening epoxy resins. Effect of Molecular Structure. Table III shows the effects of the molecular structure of the liquid polymer on the fracture energy of toughened systems. The CTIN is a carboxyl terminated isoprene-acrylonitrile copolymer; CTBS is a carboxyl terminated butadiene and styrene copolymer, and C T A is a copolymer of ethyl acrylate-butyl acrylate. We demonstrated previously that the elastic character of the rubbery domain is an important factor (4) in toughening epoxy resins. However the moduli of these liquid polymers when cured with a stoichiometric amount of epoxy resin (Epon 828) all have similar values (2.8-4.1 MPa at 300-500% elongation). Table III.

Liquid

Effect of Molecular Structure of Liquid Polymers on Fracture Energy of Epoxy Resin"

Polymer

CTB CTBN CTBNX CTIN CTBS CTA

0

Fracture Energy hJ/m

Rubbery Domain

8.8 28.8 17.5 22.8 5.3 22.8

(incompatible) yes yes yes (incompatible) yes

2

Epon 828 (100 parts), liquid rubber (5 phr), and piperidine (5 phr); cured at 120°C for 16 hr. C T (carboxyl terminated), B (butadiene), N (acrylonitrile), X (pendant-COOH group), I (isoprene), S (styrene), A (acrylate). a

6

The solubility of C T B N in epoxy resin increases directly with the acrylonitrile content (3). Systems toughened with C T B and CTBS have the least improvement in fracture energy. This probably arises from the poor solubility of the polymers in epoxy resin; however the solubility of C T A is very good. Although the particle sizes from acrylate polymers are rather small, the improvement in fracture energy is very good. The initial solubility of a reactive liquid polymer in epoxy resins before cure is a critical condition not only for chemical reactions but also for the in situ formation of rubbery domains. A good initial solubility


332

TOUGHNESS A N D BRITTLENESS

Table IV.

Effect of Level of Piperidine on Mechanical Properties

Epon 828, parts C T B N , parts Piperidine, parts Second phase: rubbery domain domain size, Mm Tensile strength, M Pa Fracture energy, hJ/m Gardner impact, J (0.635 cm thick specimen) Heat distortion temp, °C Molar ratio; piperidine/Epon 828 2


O F PLASTICS

A

B

100 0 5

100 10 5

C 100 10 7.5

D 100 10 10

65.8 1.75 5.7

Yes 1-5 58.4 33.3 7.7

Yes 1-5 54.9 40.3 6.1

Yes 1-5 54.3 49.0 7.2

80 0.11

74 0.11

73 0.17

66 0.2S

No

—

0

• Cured at 120°C for 16 hr.

with good reactivity results normally in a system with small rubbery domains. Effect of Level of Piperidine. Infrared analyses of a system cured with piperidine show a steady increase in hydroxyl content (steps II and III in Figure 4). The hydroxyl content increases markedly after the esterification is complete. The increase in the hydroxyl content lessens at a point during the etherification (steps II and IV in Figure 4). It increases again to the final hydroxyl concentration. Gel permeation chromatographic analyses are consistent with the results of infrared analyses. Thus initial molecular weight increase is minimal during the esterification; it becomes moderate during phase separation (linear etherification) and then drastic during final gelation (crosslinking etherification). Table IV shows test results of samples with various levels of piperidine and a constant amount of Hycar C T B N . The data on fracture energy in the table show conclusively an increase in crack resistance with increasing amounts of piperidine, but there is no corresponding increase in the Gardner impact. This system shows a rate sensitivity. The slow rate cleavage test for fracture energy improves although a fast rate impact test does not. Also there is a greater deterioration of mechanical properties when the additional piperidine is used. The morphology of these systems with the additional piperidine indicates that more of the particles become less distinct in the matrix. The particle boundaries are less well defined, and the fracture surface of the particles becomes similar to the matrix—i.e., the striatums in the matrix will carry over within the particle boundaries. We believe that the additional piperidine enhances linear copolymerization to give the characteristics of flexibilization to the resin. This brings the modulus of


27.

RIEW E T A L .

333



the matrix closer to that of the domain. The optimum concentration of piperidine seems to be five parts per hundred parts of resin by weight. At this concentration it gives a reasonable fracture energy with minimal sacrifice in other mechanical properties. The molar ratio of 1:10 for the amine to epoxy appears optimum for both DMP 30 and piperidine. Effect of Level of C T B N . In Table V we varied the level of C T B N at a constant amount of piperidine. At 20 parts of C T B N we find a fourfold increase in impact strength with an 11 °C loss of heat distortion temperature. This loss of thermal properties suggests that some of the C T B N flexibilizes the epoxy matrix. The morphology of these systems all shows about the same particle size. However electron micrographs of the fracture surface of the system with 20 parts C T B N show that the particles are somewhat larger and more diffuse. Effect of Bisphenol A Addition: Two-Particle Size Systems. If we modify the Hycar CTBN-epoxy resin system so as to produce a twoparticle size system, we can produce the toughest epoxy resins. We modified the CTBN-epoxy resin system by adding bisphenol A. A manyfold improvement in fracture energy is obtained when Hycar C T B N is used with bisphenol A and epoxy resin. The maximum improvement, when 5 phr Hycar C T B N is used, is at a bisphenol A concentration of 24 parts per hundred parts of epoxy resin. Typical data describing the properties of these systems are shown in Table VI. We conclude from our morphology observations and the properties listed in Table VI that at low bisphenol A content the system shows the same appearance and therefore the same properties as the C T B N epoxy system. As the bisphenol A is increased to an optimum of 24 phr, the particles decrease in size, a two-particle size population develops, and multiple failure sites appear. An example of the bisphenol A modified system is shown in Figure 2 in which the fractograph shows one family of small particles < 0.1 /xm diameter and another family of larger particles 1-5 fim diameter. Another feature shown in this micrograph are multiple Table V.

Mechanical Properties at Various Levels of C T B N "

Epon 828, parts C T B N , parts Tensile strength, M P a Elongation at break, % Modulus, GPa Fracture energy, hJ/m Gardner impact, J Heat distortion temp, °C 2

A

B

C

D

E

100 0 65.8 4.8 2.8 1.75 5.7 80

100 5 62.8 4.6 2.5 26.3 7.9 76

100 10 58.4 6.2 2.3 33.3 7.7 74

100 15 51.4 8.9 2.1 47.3 7.7 71

100 20 47.2 12.0 2.2 33.3 24.6 69

° Cured with 5 parts of piperidine at 120°C for 16 hr.


334


Table VI. Effect of Bisphenol A Concentration on Thermal—Mechanical Properties °


Amount of Bisphenol A

0 3.25 6 12 24 36 45

Fracture Energy hJ/m 2

21 16 21 33 86 37 7

Gardner Impact, J

Heat Distortion Temperature, °C

9 7 6 7 32 7 2

78 — 83 78 82 74 71

°Epon 828 (100 parts), CTBN (5 parts), piperidine (5 parts'), bisphenol A (parts); cured at 120°C for 16 hr. failure initiation sites. These multiple failure sites indicate that much more surface is created during failure. This greater surface development extends the energy absorption volume by a large amount. We believe that these features give rate insensitivity to the modified Hycar C T B N epoxy system. The two-particle size feature of this material is shown in Figure 3 also. This is an O s 0 stained microsection of the sample and shows both the large and small particles. At higher levels of bisphenol A, the particles get smaller, and the toughening properties are decreased. The thermal-mechanical properties and toughness data on this system are described in Table VII. From the data it can be seen that in the bisphenol A modified system there is no loss in thermal ( H D T ) or mechanical properties. In addition there is a further improvement in 4

Figure 2.

Electron micrograph of a bisphenol A modified CTBNtoughened epoxy resin (X 8000)



27.

RIEW E T A L .


335

Figure 3. Electron micrograph of an OsO stained microsection of a bisphenol A modified CTBN-toughened epoxy resin (X 34,000) u

fracture energy and Gardner impact strength and a significant improvement in Izod impact strength. The Izod impact test is a high rate of loading test. We also show in this table high-speed tensile tests (6.35 m/sec). The new toughened system shows an increase in both tensile strength and elongation. This high speed test shows that the bisphenol A modified systems are less flaw sensitive than the unmodified epoxy resins. Chemistry of in situ Discrete Phase Formation. We propose a reaction mechanism involving C T B N and epoxy resin in the presence of piperidine as a curing agent. We include the role of bisphenol A in this mechanism. The schematic reaction steps are summarized in Figure 4. Table VII.

Thermal—Mechanical Properties of Bisphenol A Modified CTBN—Epoxy System

Tensile strength M P a 1 Elongation at break, % [ 0.12 mm/sec Modulus, GPa I Tensile strength, M P a 1 „ , Elongation at break, % / °' Heat distortion temp, °C Fracture energy, hJ/m Gardner impact , J Izod impact, J/m of notch Nominal 0.635 cm thick samples. ft

6 b m

2

0

/

s

e

c

Unmodified Epoxy Resin System (Control)

Bisphenol A Modified CTBNEpoxy System

65.5 4.8 2.8 73.1 7.3 83 1.75 6 0.68

64.1 9.0 2.7 95.8 11.3 83 53-88 23-34 3.5

a



+

C H

(

C H

" C H

(ID H

"

O

a

Epoxy Retin (I)

*

C ^ -

H H

I Q T

B

N

[srep l ]

T

Bl.phenol A

a

r

»

( m )

C H

- " ^ < i > T - ^ > - 0 - - - H - - ^ 0

C

» "[DGEBAl-CHj -CH-CH, 4 HO-&- [CTBN].?.OH • HO-@ | " ^ O H

Sl CTBN,lo

^ ^/

Figure 4.

0

C H +

^

(VI)

CH, CH. ^H-[iX5EBA]-CH-CH,-0-^ - C - ^ ) -0-CH, .^H.jDGEBAl-CH^CH,

T E p

PRODUCT Reactions among the (I), (HI), (IV), (V) and (VI),

« " J ^ " (IVJ-CH-CH,

L>

Schematic reaction steps in bisphenol A modified CTBN-epoxy systems

CH, t - ° ^ " ^ © -

jdu)

(I)+ CH.-CH - (DOEBA]-CH-CH -0-§-[CTBN].?-0-CH .CHfDGEBA] -CH—CH,+ ^ 0 OH y OH O

< > ^

00

•g " -

a »

C


27.

RIEW E T A L .


337

S A L T F O R M A T I O N . In our typical formulations, the equivalent molar ratio of C T B N (Ephr; 0.054, 5 phr), bisphenol A (24 phr), and piperidine (5 phr) is about 1:70:20. According to the selectivity of the carboxy lie and phenolic groups (10), all the carboxylic groups and about one-third of the phenolic groups can possibly form the salts (II) and (III) in Figure 4. Infrared spectra of the mixture showed that the salt formation of bisphenol A is not very pronounced, whereas the carbonyl band (1710 cm' ) of the carboxyl group shifted completely to 1550 cm" indicating total carboxyl salt formation. Downloaded by EAST CAROLINA UNIV on August 28, 2013 | http://pubs.acs.org Publication Date: June 1, 1976 | doi: 10.1021/ba-1976-0154.ch027

1

1

ESTERIFICATION. Infrared spectra showed that the carboxylate salts of C T B N form ester groups first by opening the epoxy groups. The product should be the linear chain extended product (IV), Figure 4. Consequently while the esterification proceeds, etherification by the phenoxide ions of bisphenol A may also take place to form product (VI), Figure 4. However, the more selective carboxyl groups control the reaction so that the major product is (IV) (see Figure 4 for the various products). E T H E R I F I C A T I O N . After most of the esterification is completed, i.e., all the C T B N has reacted, the phenoxide ion controls the reaction. The diphenoxide ion of bisphenol A couples either compounds (IV) to form compound (V) or part of epoxy resin (I) to form compounds (VI). As the less selective phenol groups are consumed, active reaction among the epoxy groups takes place by the liberated piperidine from the piperidinium salts. This is the stage at which gelation begins. The gelation reduces the solubility of product ( V ) . Therefore product (V) or its aggregate would precipitate first as large particles. Product (IV) or its aggregate, the least chain extended molecules, would precipitate as small particles mainly because of a greater solubility than product ( V ) . In the CTBN-epoxy resin system without bisphenol A we believe product (IV) type may be dissolved and distributed throughout the matrix. Because the solubility parameter difference between product (IV) and the epoxy resin matrix is less than with the bisphenol A modified epoxy resin matrix, we do not observe the second small particle size in this system. G E L A T I O N ( C O M P A T I B I L I T Y ). Phase separation occurs during the incipient gelation step but without growth in particle sizes or increase in the number of particles. The particles are formed simultaneously at an early stage of gelation. Chemically, gelation is a process of crosslinking after the completion of esterification (see Figure 4). Physically, it is a process of continuously increasing the viscosity of a polymer solution with time, depending on the nature of the polymer solvent system and on the temperature. The



338


final result is solidification of the entire system into a three-dimensional network. A peculiarity of a polymer solution is that under certain conditions the system does not separate into two phases. For example if we add more than 10 phr of DMP 30 in the system (100 parts Epon 828 and 5 phr C T B N ) , the cured product is a transparent material. This occurs when the critical solubility of a polymer can be varied by adding to the solution a small quantity of a good solvent, e.g., DMP 30. Polymer Morphology and Failure Mechanisms. A failed tensile bar of unmodified piperidine-cured epoxy resin shows shear deformation before tensile failure when strained slowly (0.127 cm/sec). We could not produce stable crazes in specimens of unmodified epoxy resins. At all stress levels, temperatures, and conditions of annealing only fracture occurred after shear band formation. The failure to observe crazes in unmodified epoxy resins may be explained by a fast equilibrium condition which exists between crazing on loading and recovery on unloading. Berry (12) found a large discrepancy between observed (~ 4 X 10 ergs/cm ) and calculated (~ 10 ergs/cm ) fracture surface energy for poly (methyl methacrylate). The calculated value was computed from the energy required to rupture the chemical bonds in unit area of crosssection, assuming perfect orientation. To explain the discrepancy of fracture surface energy he proposed an orientation hypothesis of polymer molecules. He raised the possibility that a brittle crack involved plastic flow on a molecular scale at the crack tip. To produce the extent of orientation necessary to give the interference effects, a considerable amount of energy is required to orient the molecule prior to the brittle failure. Kambour (13) elucidated the nature of crazes and differentiated the crazes from cracks by observing optical properties of polycarbonate. He noted that a craze can bear a tensile load and heal under heat. The crazes in the glassy polymers studied to date are 40-60% void (14, 15, 16). In essence the crazing was seen as a process of polymer rarefaction involving the conversion of strain energy to surface free energy. He also suggested that craze is different from cold drawing which appears to involve elongation by shearing at 45° to the tensile stress with no change in material density. An electron micrograph of a fracture surface of a CTBN-toughened epoxy resin is shown in Figure 5. This C T B N is particular in that the in situ formed particles are less than 0.5 /J.m in diameter. A tensile bar of this system also shows shear deformation which indicate that the small particles have not interfered with the shear deformation characteristic of the unmodified resin. The deformation bands are nearly parallel to the planes of maximum shear stress—i.e., roughly at 45° to the principal 5

2

3

2



27.

RIEW E T A L .

Figure 5.


339

Electron micrograph of a CTBN-toughened epoxy resin with small particles (X4400)

axis of deformation. Once a deformation band forms and attains a certain width, it grows only in length. According to the electron micrograph, the deformation is a true uniform plastic shear without any sign of fracture. The morphology of a small particle size, toughened system has been further explored by examining the fracture surfaces of the stressed tensile bars. An electron photomicrograph of a replica of this surface showed a deformation band extending diagonally (Figure 6). The rubber par-

Figure 6. Electron micrograph of a CTBN-toughened epoxy resin with small particles—fracture through a shear band (X5200)



340


tides within this band are distorted in the direction of the shear deformation. The band widths are generally 5-50 /tin. However if we examine a tensile bar made from a toughened system containing large particles, we find that the material deformation arises entirely from crazing. A failed tensile bar of this system has no shear deformation or necking throughout the length of the sample. Crazing during stress is observed as whitening in the strained section of the sample. There is also a corresponding increase in the sample volume. Shear deformation does not produce whitening, and there is no increase in sample volume. The crazes occur in a direction perpendicular to the principal tensile stress axis. Both the shear bands and crazes disappear when the sample is heated above the glass transition temperature. We observed in crazed tensile samples that particles are elongated in the direction of the tensile stress field and show what appears to be a microvoid associated with each of the large particles (Figure 7). We did not observe thefibrouscrazes reported elsewhere (14,15,16) for the thermoplastics. Bucknall and Smith (17) concluded that crazing is the dominant mechanism to toughen high impact polystyrene and related polymers. One important function of the rubber particles is to serve as craze initiators and stabilizers in the glassy matrix. However Newman and Strella (18) concluded from optical microscope studies that cold drawing is responsible for toughness in ABS. The question arises as to whether our observed phenomenon which we call' rnicrovoids" is truly a microvoid associated with the particles or

Figure 7. Electron micrograph of a CTBN-toughened epoxy resin with large particles—fracture through a craze volume (X5200)


27.

RIEW E T A L .


341

an artifact produced by the fracture surface development process. We offer three reasons for proposing actual microvoids. First, the phenomenon is not observed in normal rapid cleavage surface development. Second, the craze lines extend only to the void surface and do not terminate (initiate) at the particle (Figure 7). Third, we found that a crazed tensile specimen that shows whitening (and volume increase) will shrink to its original size, and the whitening will disappear when heated above the H D T of the sample. The morphology of this heated sample showed that the microvoid evidence disappeared. Additional evidence was obtained by an O s 0 stained microsection of the stressed and heat-treated material. The particles (stained) are of the same size as in the original unstressed sample. Next we looked at the microvoid situation in a bisphenol A modified CTBN-epoxy system. This sample had the highest toughening properties that we developed in the epoxy system because of a two-particle size rubber population that uniquely gives a combination of shear deformation and tensile crazing. Only some of the large particles had microvoid development. Consequently the whitening was much less than when only crazing occurs. The multiple failure sites were still evident. Recently McGarry (19) suggested that toughness may be enhanced by the interaction of crazes and shear bands in a system with a bidistribution of particle sizes. He (20) later showed the effect of rubber particle size on deformation mechanisms in C T B N toughened epoxy system—a shear mechanism is enhanced by the presence of particles a few hundred Angstroms in diameter; whereas crazing appears as the predominant flow mechanism under tensile stress with larger particles (1.5-5 /*m diameter). Bucknall et at (21, 22) supported McGarry's observations for shear bands and crazes in rubber modified polyphenylene oxide. He found that HIPS/PPO blends deformed by a combination of crazing and shear band formation; whereas the HIPS deformed almost entirely by crazing. The results in Table VII show conclusively that there is a synergistic benefit gained in the two-particle size system in toughened properties. This benefit arises from the dual energy-absorbing mechanisms of combined shear deformation and crazing.


4

Conclusions We conclude that: (a) Shear deformations are dominant in resins toughened with small rubbery particles, (b) Crazing associated with polymer whitening and microvoid development is dominant in resins toughened with large particles, (c) Maximum toughening is obtained under conditions of combined shear and craze deformations. This condition is obtainable when both large and small particles are present.



342


(d) Conditions of maximum toughening also show multiple failure sites in fractured surfaces. These sites allow for a greatly increased surface development, and (e) from the present study and our previous findings a toughened system must meet the following criteria: Chemical Criteria. (a) The liquid polymer should have at least two functional end-groups, (b) The functional group must have a selectivity to form a salt with an amine cure agent. Also it should have a reactivity which enables it to undergo initial linear chain extension with a difunctional epoxy resin. The selectivity and reactivity of the functional group are necessary to form chemical bonds between the rubbery particles and the epoxy matrix, (c) The liquid polymers should have some degree of solubility in the epoxy resin without phase separation or agglomeration before gelation. Morphological Criteria. It must show a second particulate rubbery phase. The second phase must have a definite size and shape and the domains must be dispersed throughout the matrix. Thermal-Mechanical Criteria. The thermal and mechanical properties must not be degraded. The toughness properties (crack and impact resistance) must be improved, and the system should be rather insensitive to the rate of loading (or temperature) of a given test. Acknowledgments The authors thank the B. F. Goodrich Co. for permission to publish this paper. Literature

Cited

1. McGarry, F. J., Willner, A. M., "Toughening of an Epoxy Resins by an Elastomeric Second Phase," Res. Rept. R68-8, School of Engineering, Massachusetts Institute of Technology, Cambridge, Mass. (1968). 2. Sultan, J. N., McGarry, F. J., "Microstructural Characteristics of Toughened Thermoset Polymers," Ibid. R69-59 (1969). 3. Rowe, E. H., Siebert, A. R., Drake, R. S., Mod. Plast. (1970) 417, 110. 4. Siebert, A. R., Riew, C. K., "The Chemistry of Rubber Toughened Epoxy Resins I," 161st National Meetg., ACS, Org. Coatings Plast. Div. (March 1971). 5. Rowe, E. H., "Fractography of Thermoset Resins Toughened with Hycar CTBN," 26th Ann. Tech. Conf., Reinforced Plast. Composites Inst., SPI, Section 2-D (1971). 6. Riew, C. K., Smith, R. W., J. Polym. Sci., Part A-1 (1971) 9, 2737. 7. Rowe, E. H., Riew, C. K., "What Failure Mechanisms Tell about Toughened Epoxy Resins," Plast. Eng. (March 1975) 45-47. 8. Berry, J. P., "Determination of Fracture Surface Energy by a Cleavage Technique," American Physical Society Meeting, Baltimore, Md. (April 1962). 9. Andrews, E. H., Walsh, A., Nature (1957) 179, 729.



27. RIEW ET AL. Rubber Toughened Thermosets

343

10. Riew, C. K., Rowe, E. H., Siebert, A. R., Lipiec, J. M., "ABE Thermosets for Structural Plastics," 27th Ann. Tech. Conf., Reinforced Plast. Composites Inst., SPI, Section 2-D (1972). 11. Siebert, A. R., Rowe, E. H., Riew, C. K., Lipiec, J. M., "Toughness vs. Flexibility in Epoxy Resins," 28th Ann. Tech. Conf., Reinforced Plast. Composites Inst., SPI, Section 1-A (1973). 12. Berry, J. P., Nature (1960) 185, 91. 13. Kambour, R. P., Nature (1962) 195, 1299. 14. Kambour, R. P., Polymer (1964) 5, 143. 15. Kambour, R. P., J. Polym. Sci., Part A (1964) 2, 4159. 16. Kambour, R. P., Polym. Eng. Sci. (1968) 8, 281. 17. Bucknall, C. B., Smith, R. R., Polymer (1965) 6, 437. 18. Newman, S., Strella, S.,J.Appl.Polym. Sci. (1965) 9, 2297. 19. McGarry, F. J., Proc. Roy.Soc.,A (1970) 319, 59. 20. Sultan, J. N., McGarry, F.J.,Polym. Eng. Sci. (1973) 13, 29. 21. Bucknall, C. B., Clayton, D., Keast, W. E., J. Mater. Sci. (1972) 7, 1443. 22. Ibid. (1973) 8, 514. RECEIVED

October 18, 1974.


Rubber Toughened Thermosets - Advances in Chemistry (ACS

Recommend Documents